The most underrated research skill isn’t stats or coding, it’s asking “what would I expect to see *if I were wrong*?”—without that, you’re not analyzing reality, you’re just running victory laps around your own assumptions. #datamindset